Co-crystal formation with 1,2,3,5-dithiadiazolyl radicals

Article abstract of DOI:10.1039/c3ce41519h

The crystal structure of a novel dithiadiazolyl co-crystal, the second such co-crystal to be reported, has been determined, and is described in the context of the structures of the co-crystal formers. It is also shown that both known dithiadiazolyl co-crystals ([PhCN₂S₂][C₆F ₅CN₂S₂] and [PhCN₂S ₂][NC₅F₄CN₂S₂]) can be made via co-sublimation or by mixing co-crystal formers in solution, but not mechanochemically. DSC and theoretical calculations reveal that the thermodynamics of co-crystal formation in these systems is very finely balanced.

Full text of DOI:10.1039/c3ce41519h

CrystEngComm
View Article Online
View Journal
PAPER
Co-crystal formation with 1,2,3,5-dithiadiazolyl
radicals†
Cite this: DOI: 10.1039/c3ce41519h
a
a
b
Sean W. Robinson, Delia A. Haynes* and Jeremy M. Rawson
Received 31st July 2013,
The crystal structure of a novel dithiadiazolyl co-crystal, the second such co-crystal to be reported, has been deter-
mined, and is described in the context of the structures of the co-crystal formers. It is also shown that both known
Accepted 16th October 2013
2 2 6 5 2 2 2 2 5 4 2 2
dithiadiazolyl co-crystals ([PhCN S ][C F CN S ] and [PhCN S ][NC F CN S ]) can be made via co-sublimation or by
DOI: 10.1039/c3ce41519h
mixing co-crystal formers in solution, but not mechanochemically. DSC and theoretical calculations reveal that the
thermodynamics of co-crystal formation in these systems is very finely balanced.
www.rsc.org/crystengcomm
−
1
Introduction
reveal the dimerisation enthalpy in PhCN S (ca. −35 kJ mol )
2
2
7
is large, although EPR studies have recently suggested smaller
dimerisation enthalpies for other derivatives in the solid state
~ −10 to −20 kJ mol ). Nevertheless, these interactions are
thermodynamically significant on the scale of intermolecular
interactions and it is, perhaps, unsurprising that the majority
of DTDA radicals dimerise in the solid state. Inhibition of
this dimerisation process must be overcome if any interesting
magnetic or conducting properties are to be observed, and
previous studies have reflected some modest success in using
Dithiadiazolyl radicals have been the focus of a number of
studies investigating their potential as magnetic or con-
ducting materials. The 1,2,3,5-dithiadiazolyl isomer in par-
ticular (hereafter referred to as DTDA) has received much
attention, due in part to the well-established synthetic routes
to a variety of derivatives, and to the thermal and kinetic sta-
−1
8
(
1
bility of these radicals. The observation of weak ferromag-
2
netic ordering at 36 K in the β-phase of 4′-NCC
well as ferromagnetic ordering at 1.3 K in p-O NC
has led to a great deal of interest in this family of radicals.
DTDAs are excellent building blocks for organic magnetic
or conducting materials: the unpaired electron is localised
on the heterocyclic ring in a SOMO of a symmetry which is
nodal at the carbon atom. This in turn implies that
functionalisation at the R-group does not significantly affect
the electronic properties of the radical-containing heterocycle
6
F
4
CN
2
S
2
˙, as
3
2
6
F
4
CN
2
S
2
˙,
2,3
9
perfluorophenyl substituents and steric effects to overcome
dimerisation, whilst chlorophenyl substituents have been shown
8
to weaken the dimerisation interaction in the solid state.
We have recently begun investigating the potential of
co-crystallisation as a means of overcoming dimerisation
in DTDA radicals. This is in part based upon: (i) our
increasing understanding of the structure-directing inter-
actions associated with these radicals and the potential to test
our ability to generate pre-defined molecular architectures;
and (ii) the potential for modulation of the electronic or
magnetic behaviour of these materials through introduction
of a co-crystal former. Several radical-containing co-crystals
have been reported in the literature, and the strategy clearly
shows promise for developing materials with interesting
magnetic properties.
Recently we reported the synthesis and crystal structure
of the first DTDA-DTDA co-crystal, [PhCN ][C CN ] (4)
from 1 and 2 (Scheme 1). We now report the synthesis and
characterisation of second DTDA-DTDA co-crystal,
[PhCN ][NC CN ] (5) from 1 and 3, as well as extensive
2
4
(
which has been confirmed through EPR and electrochemical
5
studies). This makes DTDA radicals interesting to study
from a crystal engineering perspective, and the effect of the
R-group on the solid-state structures of these materials has
6
been investigated by several groups.
On the other hand, DTDAs tend to dimerise in the solid
state, rendering them diamagnetic insulators. Solution studies
10
1
1
a
Department of Chemistry and Polymer Science, University of Stellenbosch,
2
S
2
6
F
5
2 2
S
P. Bag X1, Matieland, 7602, South Africa. E-mail: dhaynes@sun.ac.za
b
Department of Chemistry and Biochemistry, University of Windsor,
a
401 Sunset Avenue, Windsor, ON, Canada, N9B 3P4
†
Electronic supplementary information (ESI) available: PXRD patterns for 1–5,
2
S
2
5
F
4
2 2
S
list of attempted co-crystal syntheses, DSC data, thermodynamic cycle for for-
mation of 5. CCDC 952623 contains the supplementary crystallographic data
for this paper. These data can be obtained free of charge from The Cambridge
Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif. For ESI
and crystallographic data in CIF or other electronic format see DOI: 10.1039/
c3ce41519h
characterisation of the two co-crystals. Various preparative
routes to obtain co-crystals 4 and 5 are also reported, and the
thermodynamics of co-crystal formation probed through DSC
studies. A large number of attempts to produce further
co-crystals containing DTDA radicals were unsuccessful,
This journal is © The Royal Society of Chemistry 2013
CrystEngComm

View Article Online
Paper
CrystEngComm
obtained as red blocks in yields ranging from 50–61%
(
based on mass of crude salt).
+
(
+)-ESI-MS: m/z 270.9 (M , 100%)
EPR (298 K, CH Cl ): quintet (g = 2.010, a = 4.9 G).
2
2
N
Synthesis of 4-perfluoropyridyl-1,2,3,5-dithiadiazolyl, 3. A
Scheme 1
water-cooled cold finger was used in the purification of 3,
and solids were heated to 100 °C under vacuum. Crystals of 3
were obtained as lustrous blue blocks. Yield = 19.0% based
on hexamethyldisilazane.
reflecting the subtle energy balance in co-crystal formation in
these systems. These experiments are also described herein.
+
(+)-ESI-MS: m/z 253.9 (M , 30%), 221.1 (16), 194.0 (100)
EPR (298 K, CH Cl ): quintet (g = 2.011, a = 5.0 G).
2
2
N
PXRD patterns of 1–3 matched the simulated patterns
Experimental
1
1,13
based on reported single crystal data.
These are given in
the ESI.†
Chemicals were purchased from Sigma-Aldrich, and used as
received. Radicals 1–3 were prepared using standard Schlenk
techniques, following a procedure modified from that reported
in the literature. The synthesis of 1 is described as a typical
procedure; 2 and 3 were prepared in an analogous fashion.
Synthesis of co-crystal 4. A 1 : 1 mole ratio of 1 and 2 was
placed in a sublimation tube. The solids were sublimed
under vacuum at 65 °C to yield crystals of 4 as red blocks,
whose structure and phase purity were determined by single
crystal and powder X-ray diffraction.
12
2 2 N
EPR (298 K, CH Cl ): quintet g = 2.010, a = 5.0 G.
Synthetic procedures
Synthesis of co-crystal 5. A 1 : 1 mole ratio of 1 and 3 was
placed in a sublimation tube. The solids were sublimed
under vacuum at 80 °C to yield crystals of 4 as red blocks,
whose structure and phase purity were determined by single
crystal and powder X-ray diffraction.
Synthesis of 4-phenyl-1,2,3,5-dithiadiazolyl, 1. A Schlenk
tube containing 1,1,1,3,3,3-hexamethyldisilazane (1.20 ml,
5
.69 mmol) in diethyl ether (20 ml) was cooled to −78 °C
using a dry ice–acetone bath. A solution of n-butyllithium in
hexanes (3.80 ml, 6.08 mmol) was added slowly, after which
the dry ice–acetone bath was removed, resulting in a milky
solution which became clear on reaching room temperature.
Benzonitrile (0.60 ml, 5.82 mmol) was added, yielding a clear
pale yellow solution which was stirred overnight. The
2 2 N
EPR (298 K, CH Cl ): quintet g = 2.010, a = 4.9 G.
For both 4 and 5 the two components could not be
resolved in the EPR spectrum due to the close similarity in
1
4
g-values and N hyperfine coupling constants for 1, 2 and 3.
PXRD patterns for 4 and 5 are discussed later.
solution was cooled to 0 °C, and SCl (1.00 ml, 15.74 mmol)
2
X-ray diffraction
was added dropwise, yielding an immediate bright yellow
precipitate. The solution was allowed to warm to room
temperature, and stirred for five hours, after which the
resultant bright yellow solid was filtered, washed with diethyl
ether (2 × 15 ml) and dried in vacuo.
Solid triphenylantimony (0.788 g, 2.23 mmol) was added
directly to the crude salt. The mixture was heated to 70 °C
under a nitrogen atmosphere until the colour change to
purple was complete. A water-cooled cold finger was placed
in the Schlenk, and the solids were heated to 90 °C under
vacuum, affording green-black needles of 1 (0.287 g, 1.59 mmol,
Single crystal X-ray diffraction data for 5 were collected on a
Bruker MicroSTAR-H diffractometer with an ApexII detector
using monochromated Cu-Kα radiation. The crystal was
mounted in Paratone-N oil and cooled to 100 K using an
Oxford Cryosystems Cryostream 700 cryostat. Data reduction,
absorption corrections and unit cell determination were car-
ried out using the diffractometer software (APEXII, Bruker).
Structures were solved and refined using the SHELX-97
package implemented through X-Seed. Hydrogen atoms
were placed in calculated positions using riding models.
1
4
1
5
1
6
1
7
2
7.9% yield).
Figures were generated using Mercury.
Several samples were prepared in this way. Representative
Powder X-ray diffraction patterns were collected on a
PANalytical X'Pert PRO diffractometer in Bragg–Brentano
geometry using an X'Celerator detector and Cu radiation
source with a Ni filter. Samples were spun during data collec-
tion. Powder patterns were simulated and visualised in X'Pert
analytical data are given below.
+
+
(
+)-ESI-MS: m/z 180.9 (M , 85%), 135.0 (M − SN, 38),
+
7
7.9 (SN , 100)
2
2 2 N
EPR (298 K, CH Cl ): quintet (g = 2.007, a = 5.0 G).
Radicals 2 and 3 were prepared in an analogous fashion
to 1 and provided satisfactory PXRD profiles, MS and
EPR spectra.
1
8
HighScore Plus. In all powder diffraction patterns, the
y-axis gives intensity counts in arbitrary units.
Differential Scanning Calorimetry (DSC) was carried out
Synthesis of 4-perfluorophenyl-1,2,3,5-dithiadiazolyl, 2. A
cold finger containing a mixture of dry ice and acetone was
used in the purification of crude 2. Several preparations were
carried out, with solids sublimed under vacuum at
temperatures between 70 and 120 °C. Crystals of 2 were
using a TA Instruments Q100 system under a N gas
2
−1
purge, with a flow rate of 50.0 ml min . The ramp rate was
−
1
−1
10 °C min , and the cooling rate was 5 °C min . Samples
were placed in aluminium pans that were non-hermetically
sealed with non-vented aluminium lids.
CrystEngComm
This journal is © The Royal Society of Chemistry 2013

View Article Online
CrystEngComm
Paper
Calculations
Single-point energy DFT calculations were undertaken on
the dimer structures determined by X-ray diffraction. All
computations were undertaken at the B3LYP/6-31G*+ level
1
9
with a closed shell singlet configuration within Jaguar.
Thermodynamic calculations were zero point energy corrected
and determined at 298.15 K. Notably the structures of 1–5 were
not energy minima on the potential energy surface, as reflected
in several negative vibrational frequencies in all cases. However
geometry optimisation was not attempted since this would not
accurately reflect the energetics associated with the crystal
structure geometries.
Fig. 1 The two crystallographically independent dimers in the asymmetric unit of 5.
Results and discussion
The preparation of crystals of 4 and 5 suitable for single-
crystal X-ray diffraction was achieved by co-sublimation. The
crystal structure of the new co-crystal, 5, will be described,
followed by investigations into alternative methods of prepar-
ing both 4 and 5. Thermodynamic studies on the formation
of 4 and 5 will also be discussed, as well as the implications
of these investigations for other co-crystallisation attempts.
Crystal structure of 5
Crystals of 5 were grown by sublimation as blue-purple
blocks, which crystallise in the orthorhombic space group
Pcab.‡ The asymmetric unit contains two heteromeric cis-oid
dimers, i.e. each dimer consists of one molecule of [PhCN S ]
2
2
5 4 2 2
and one molecule of [NC F CN S ] (Fig. 1). The intra-dimer
S⋯S distances (2.9574(5)–3.0676(6) Å) are comparable
with other DTDA dimers e.g. 1 (3.016(1)–3.147(2) Å) and
pyr
Fig. 2 Packing in 5 and 3. (a) Chains of dimers formed via S–S⋯N interactions.
1
1
2
(3.111(2) Å). The twist angles between the DTDA and
(
b) Sheets of dimers in 5 as viewed along the crystallographic c axis. (c) Sheets of
aryl/pyridyl are somewhat varied. The two DTDAs with
fluoro-pyridyl substituents exhibit large twist angles (26.52 and
dimers in 3 as viewed along the crystallographic c axis. Note the similarity to (b).
3
0.95°), albeit at the lower end of those reported for DTDA
2
0
radicals with two ortho-fluorine substituents (24.9–68.7°).
(
3.241(1)–3.480(1) Å, Fig. 2a). Unlike other S–S⋯N contacts
Conversely whilst one phenyl-substituted DTDA exhibits a
such as S–S⋯NC, the pyridyl-N atom is displaced from the
DTDA ring plane, presumably so as to optimise additional
S–S⋯N contacts to the second NC F CN S molecule in each
small twist angle (6.02°) comparable with other aryl derivatives
20
with two ortho-H (5.0–11.8°), the other is markedly larger
26.39°). DFT calculations on both fluorophenyl and phenyl
5
4
2 2
(
dimer. These longer S–S⋯N contacts span the range
derivatives reveal shallow energy minima at 50° and 0° respec-
tively but with ranges of 30–90° and 0–25° all accessible with a
3
.669(1)–3.765(1) Å. Chains are linked into sheets in the
ab plane via side-on S⋯N interactions (3.295(1)–3.504(1) Å).
Sheets stack along c in an antiparallel fashion.
The most significant feature of the structure of 5 is the
supramolecular synthons that are also seen in the structure
of pure 3 – the S–S⋯N
(
−1
20
modest energy cost (~3 kJ mol ), suggesting that the overall
molecular geometries in 5 are not significantly strained.
Dimers are linked into chains along the crystallographic
b-axis via contacts between the electropositive S–S of the phenyl
radical in one dimer and the pyridyl nitrogen of another dimer
1
3
pyr
and side-on S⋯N contacts
Fig. 2c). This implies that synthons from the crystal struc-
ture of a co-crystal former may also be significant in the
structure of the co-crystal. In fact, the crystal structures of 3
and 5 are near superimposable. A closer examination of the
unit cell parameters reveals that, whilst 3 and 5 are not
isostructural, these two structures are very closely related; 3
adopts the monoclinic setting I2/a (an alternative setting of
C2/c, no. 15) with a = 11.420(4), b = 9.654(4) and c = 14.563(4) Å
and β = 94.06(1). Co-crystal 5 crystallises in the space group
‡
0
1
3
2
5 4 5 4
Crystal data for 5: C13H F N S , M = 435.46, blue-purple block, 0.09 × 0.06 ×
3
.06 mm , orthorhombic, space group Pcab (no. 61), a = 10.8996(4), b =
3
−3
9.1551(6), c = 29.1067(10) Å, V = 6077.0(4) Å , Z = 16, D
c
= 1.904 g cm , F000
=
488, Bruker APEX-II CCD, CuKα radiation, λ = 1.54178 Å, T = 173(2)K,
max = 136.5°, 32 208 reflections collected, 5530 unique (Rint = 0.0179). Final
= 0.0237, wR = 0.0658, R indices based on 5378 reflections
θ
GooF = 1.096, R
with I > 2σ(I) (refinement on F ), 469 parameters, 0 restraints. Lp and absorption
1
2
2
−1
corrections applied, μ = 6.295 mm . Unconventional space group setting chosen
to facilitate comparison with 3.
This journal is © The Royal Society of Chemistry 2013
CrystEngComm

View Article Online
Paper
CrystEngComm
Pcab with a = 10.8996(4), b = 19.155(6) and c = 29.1067(10) Å.
The b-axis of 5 is approximately double the b-axis of 3, and the
c-axis of 5 is approximately double that of 3. The doubling of
two unit cell axes, as well as the occurrence of heterodimers in
thermodynamic balance is anticipated, in which the strength
of inter-dimer interactions will play a large part in determin-
ing the overall favourability of the reaction (Table 2).
5, in combination with the change from a body-centred to a
Preparation of 4 and 5 by sublimation
primitive cell, is entirely consistent with the increase of num-
ber of molecules in the asymmetric unit from one molecule in
the asymmetric unit of 3 to four molecules in 5. In terms of
using co-crystal formers to control crystal structure this result
is very promising – it seems that strong structure-directing
interactions of one co-crystal former can be implemented to
direct the structure of the other co-crystal former.
Crystals of both
4 and 5 were initially prepared by
co-sublimation of a 1 : 1 mixture of the respective co-crystal
formers. In order to determine whether only co-crystal was
produced by this method, or whether a mixture of co-crystal
and co-crystal formers was obtained, the following experi-
ment was carried out. A 1 : 1 mixture of the co-crystal formers
was placed in a sublimation tube, and the solids heated
under vacuum until all material had sublimed. The sublimed
crystals were then scraped out of the sublimation tube, and a
PXRD pattern collected. In both cases (1 + 2 and 1 + 3), the
product was shown to be pure co-crystal (within the limits of
detection of PXRD) (Fig. 3).
Attempted synthesis of further co-crystals
The successful synthesis of co-crystals 4 and 5 prompted the
investigation of co-crystal formation with a variety of combi-
nations of other DTDA radicals, as well as combinations of
DTDA radicals with various small organic molecules. In all
cases, co-sublimation of a 1 : 1 molar ratio of co-crystal
formers was used, except if one of the two components was a
liquid, in which case one component was dissolved in the
other. A full list of combinations of molecules used is given
in the ESI.† In all cases, either crystals of the co-crystal
formers were obtained, or no crystals were obtained.
In both cases, co-reducing the two relevant dithia-
diazolylium salts, followed by sublimation, also gave pure
co-crystal.
Table 2 Calculated thermodynamic parameters for dimers of 1, 3 and co-
crystal 5 (energies in hartrees unless otherwise stated)
Compound
[PhCN S ]
H
G
TS
These results raised the question of why co-crystals 4 and
5
appear to form so readily. Similarity in sublimation tempe-
−2350.819891 −2350.883403 0.063512
−3177.015490 −3177.096345 0.080855
−2763.921451 −2763.993011 0.071560
2
2 2
[
[
NC
PhCN
5
F
4
CN
2 2 2
S ]
ratures of the two components does not appear to be a
significant feature: the sublimation temperatures of 1, 2 and 3
are approximately 100, 50 and 110 °C respectively. Theoretical
calculations and thermodynamic studies were therefore carried
out to shed some light on the stability of co-crystals 4 and 5.
2
S
2
][NC
5
F
4
2
CN S
2
]
Δ(rxn) (hartree)
Δ(rxn) (kJ mol
−0.007521
−19.7
−0.006274
−16.5
−0.001247
−3.3
−
1
)
Theoretical investigation of 4 and 5
Single-point energy calculations of the energies of dimers of
1, 2 and 4 reveal that the reaction (eqn (1)) is favoured in the
forward direction, albeit by a small amount (Table 1).
½
PhCN2S2ꢀ þ ½C6F5CN2S2ꢀ → 2½PhCN2S2ꢀ½C6F5CN2S2ꢀ ð1Þ
2
2
Assuming that the entropy and enthalpy are not signifi-
cantly temperature dependent, this reaction should be
exergonic across all temperatures, since both enthalpy and
entropy terms appear favourable.
When the same analysis is applied to the reaction of 1
with 3, the reaction is found to be enthalpically favourable
but marginally entropically disfavoured and a much finer
Table
1 Calculated thermodynamic parameters for dimers of 1, 2 and
co-crystal 4 (energies in hartrees unless otherwise stated)
Compound
H
G
TS
[
[
[
PhCN
CN
2 2 6 5 2 2
PhCN S ][C F CN S ]
2
S
S ]
2 2 2
2
]
2
−2350.819891
−3343.365223
−2847.094468
−0.0038222
−10.0
−2350.883403
−3343.449343
−2847.170952
−0.009158
0.063512
0.084120
0.076484
0.005336
14.0
Fig. 3 Preparation of co-crystals 4 and 5 by sublimation. (a) Preparation of 4 and
b) preparation of 5. In both cases PXRD shows that the product obtained is pure
6
C F
5
(
Δ(rxn) (hartree)
Δ(rxn) (kJ mol
co-crystal. 5 has some residual peaks from 1, but none from 3. Presumably this is
due to a very slight excess of 1 in the co-sublimed mixture.
−1
)
−24.0
CrystEngComm
This journal is © The Royal Society of Chemistry 2013

View Article Online
CrystEngComm
Paper
Table 3 Melting and recrystallisation temperatures of 1–5 as determined
Attempts to produce 4 and 5 mechanochemically
by DSC
Similar experiments were carried out to determine whether 4
and 5 could be produced mechanochemically. A 1 : 1 mixture
of the relevant co-crystal formers was ground by hand in a
dry mortar and pestle for 30 seconds in air (grinding for any
longer resulted in one of the co-crystal formers melting). In
both cases (1 + 2 and 1 + 3), the product obtained was a phys-
ical mixture of the co-crystal formers (see ESI†).
Onset
temperature
Peak
maximum
Peak maximum
(recrystallisation) (kJ mol
ΔHfus
−1
(melting) (°C) (melting) (°C) (°C)
of dimer)
1
2
3
4
5
116.0
42.0
134.2
100.3
98.4
119.8
45.0
136.0
102.2
101.0
87.9
Not observed
125.3
61.4
74.9
57.6
36.1
57.7
61.7
54.6
a
a
Solid state phase transition observed at ~80 °C.
Preparation of 4 and 5 from solution
Mixing a 1 : 1 ratio of co-crystal formers in a small volume of
THF, followed by removal of the solvent in vacuo, yielded
co-crystals 4 and 5 (Fig. 4). It is known that dithiadiazolyl
radicals exist in equilibrium between the monomer and the
dimer in solution. Presumably this equilibrium allows for
the dissociation of dimers of the co-crystal formers, and
preferential formation of co-crystal heterodimers.
−1
1
50 °C with a heating rate of 10 °C min , then finally cooled
−1
to room temperature at 5 °C min . Results are summarised
in Table 3.
DTDA 3 showed evidence of a phase change just above
0 °C. Initial VT-PXRD experiments confirm that this is a
structural change. Investigations are currently underway to
further characterise this potential new polymorph, and will
be the subject of a future report.
7
8
Preparation of 4 and 5 from the melt
Initial attempts to form co-crystals 4 and 5 from the melt of
Table 3 shows that the melting points of the co-crystals
are distinct from the melting points of the pure co-crystal for-
mers. Melting point could therefore potentially be used to
establish if co-crystals form in the melt. Experiments were
carried out with a 1 : 1 mixture of co-crystal formers in a
sealed DSC pan. These were cooled, heated and re-cooled as
described above.
1
+ 2 or 1 + 3 gave very small yields of co-crystal, largely due
to sublimation of one of the co-crystal formers in both cases
1 + 3, where the co-crystal formers sublime at higher tempe-
(
ratures, gave better results). The use of differential scanning
calorimetry to study the formation of these co-crystals was
therefore investigated.
The melting and recystallisation temperatures of 1–5 were
determined in sealed pans (to prevent sublimation). Samples
were cooled to −80 °C at a rate of 5 °C min , then heated to
Previous studies by Passmore and coworkers have revealed
that there is a marked increase in both molecular volume
and paramagnetism upon melting of thiazyl radicals, consis-
tent with a breakdown of the π*–π* dimer upon formation of
the liquid phase. Whilst complete dissociation of dimers in
the liquid phase is a dangerous assumption, the enthalpies
of fusion of 1, 2 and 4 can be used to provide an initial esti-
−
1
2
1
mate of the thermodynamic favourability of the reaction 1
2
(s) +
2
2
(s) → 2 4(s). From the DSC data in Table 3, a small favourable
−1
reaction enthalpy is estimated (ΔH
= −30 kJ mol or
rxn
−1
−
15 kJ mol of 4) for this process. This approach, however,
ignores any mixing enthalpy associated with combining 1 and
in the melt phase. This is likely to be small, and has been
2
directly estimated using the approach outlined in Fig. 5, which
utilises data from in situ reaction of 1 with 2.
DSC studies on a 1 : 1 mixture of 1 and 2 (Fig. 6a) reveal
an endotherm at 41–46 °C, with ΔH = 32 kJ per mole of dimer
2
. Both the position and ΔH value for this endotherm are
consistent with the melting of 2 (Table 3). This was followed
by another endotherm at 96–100 °C. Whilst this endotherm
occurs near the melting point of pure 4 (Table 3), the
Fig. 4 Preparation of co-crystals 4 and 5 from solution. (a) Preparation of 4 and
(
b) preparation of 5. In both cases, PXRD shows that the product obtained is pure
Fig. 5 Thermodynamic estimate of ΔHrxn for formation of co-crystal 4 from (1)
and (2)
2
co-crystal.
2
.
This journal is © The Royal Society of Chemistry 2013
CrystEngComm

View Article Online
Paper
CrystEngComm
some radical mobility occurs prior to melting (mp 98–101 °C).
This is not unexpected as most DTDA radicals are purified
by sublimation, i.e. they have mobility below their melting
point. The total enthalpy change associated with these
−
1
transitions (96 kJ mol ) is a little less than the anticipated
−
1
1
14 kJ mol for melting of 1 and the second phase of 3
−1
independently, suggesting a favourable ΔHmix of −20 kJ mol .
Taking into account the enthalpy of fusion of 5, the net
2 2
formation of two moles of 5(s) from 1 (s) and 3 (s) is exo-
−1
thermic by just 3 kJ mol (see ESI†).
Notably the enthalpy changes for formation of two
co-crystal dimers from the respective homodimers are small
−
1
−1
(
−37 kJ mol and −3 kJ mol respectively). These values are
the same order of magnitude as the computed values for the
gas-phase formation of the heterodimers (−10 kJ mol and
20 kJ mol
formation contributes significantly to the favourability of
the reaction.
−
1
−
1
−
respectively), suggesting that heterodimer
Conclusions
A
second dithidiazolyl co-crystal, containing [PhCN S ]
2 2
Fig. 6 DSC traces of a 1 : 1 mixture of (a) 1 + 2 and (b) 1 + 3.
5 4 2 2
[NC F CN S ] heterodimers, has been prepared and structu-
rally characterised. It has been shown that both known DTDA-
DTDA co-crystals can be prepared by co-sublimation of
co-crystal formers, by mixing co-crystal formers in solution or
from the melt but could not be produced mechanochemically.
These alternative synthetic routes open up the possibility
of preparing DTDA co-crystals with materials that do not
readily sublime.
Thermal analysis data clearly reveal that whilst co-crystal
formation is favoured, the enthalpy change in both cases is
only marginally favourable. The clearly fine thermodynamic
balance for co-crystal formation is evidenced by the large
number of unsuccessful co-crystallisation attempts using
different derivatives, and both paramagnetic and diamagnetic
co-crystal formers chosen for their potential to form favourable
intermolecular interactions with one another (see ESI†).
Given the small energies involved, a choice of co-crystal
formers with low enthalpies of fusion, coupled with co-crystal
products which are anticipated to exhibit particularly
strong intermolecular forces, appear promising candidates for
future study.
−
1
enthalpy change (ΔH = 50 kJ mol for two dimers of 4) is
much too low for it to be attributed to melting of 4 (62 kJ per
mole of dimer). Moreover, an additional feature in the DSC
would be anticipated for formation of solid 4 from molten 2
and solid 1 prior to this endotherm but this is not observed.
This second endotherm is tentatively attributed to simulta-
neous generation and melting of 4. From a thermodynamic
perspective, this equates to melting of 1 coupled with the
enthalpy of mixing of liquid 1 with 2 (Fig. 5). Since the
enthalpy of fusion of 1 is 58 kJ per mole of dimer then we
−
1
estimate ΔHmix as −8 kJ mol .
For an ideal mixture solvent–solvent and solute–solute
interactions are identical and ΔHmix = 0, so the current data
reflect a small negative deviation from Raoult's Law. Using
all the thermodynamic data we find ΔHrxn for formation of
from 1 and 2 is modestly enthalpically favourable (−37 kJ
for formation of two moles of 4, or ca. −18 kJ mol ). This is
comparable with the calculated enthalpy change for formation
of the heterodimer in the gas phase from the DFT studies
4
−
1
−1
(
−10 kJ mol ). This would suggest that the formation of
the heterodimer contributes significantly to the overall
favourability of the reaction.
Acknowledgements
D. A. H. acknowledges the National Research Foundation§
and Stellenbosch University for funding. S. W. R. thanks the
Harry Crossley Foundation, the NRF, the Agar Hamilton
Trust and Stellenbosch University for financial assistance.
J. M. R. acknowledges the support of the Canada Research
Chair Program, CFI and ORF. Single crystal X-ray diffraction
Using a similar approach, and taking into account the
−
1
2
enthalpy of the phase change for 3 (12.8 kJ mol ), we
estimate formation of liquid 1 and 3 from 1 (s) and 3 (s) to be
2
2
−
1
+
128 kJ mol (Table 3) which nearly compensates the forma-
tion of liquid 1 and 3 from fusion of co-crystal 5 (110 kJ per
two moles of 5). A detailed thermal analysis of the reaction
mixture of 1 and 3 showed the phase transition of 3 at
2 °C, followed by a second endotherm at 102 °C associated
with simultaneous formation and melting of 5. Since this
occurs below the melting points of both 1 and 3, we assume that
§
This material is based upon work supported financially by the National
8
Research Foundation of South Africa. Any opinion, findings and conclusions
or recommendations expressed in this material are those of the authors and
therefore the NRF does not accept any liability in regard thereto.
CrystEngComm
This journal is © The Royal Society of Chemistry 2013

View Article Online
CrystEngComm
Paper
data for 5 were collected at the X-ray crystallographic facility
at the College of Chemistry, University of California, Berkeley.
D. A. H. is also grateful for an award from the Fulbright
Scholar Program, hosted in the labs of Prof J. Long at UC
Berkeley, which enabled initial investigations in this area.
A. Alberola, E. Carter, C. P. Constantinides, D. J. Eisler,
D. M. Murphy and J. M. Rawson, Chem. Commun.,
2011, 2532–2534.
9 A. Alberola, C. S. Clarke, D. A. Haynes, S. I. Pascu and
J. M. Rawson, Chem. Commun., 2005, 4726–4728.
0 H. Akpinar, J. T. Mague and P. M. Lahti, CrystEngComm,
1
2
013, 15, 831–835; H. Akpinar, J. T. Mague, M. A. Novak,
Notes and references
J. R. Friedman and P. M. Lahti, CrystEngComm, 2012, 14,
1515–1526; G. M. Espallargas, A. Recuenco, F. M. Romero,
L. Brammer and S. Libri, CrystEngComm, 2012, 14,
6381–6383.
1
J. M. Rawson, A. Alberola and A. Whalley, J. Mater. Chem.,
006, 16, 2560–2575.
2
2
A. J. Banister, N. Bricklebank, W. Clegg, M. R. J. Elsegood,
C. I Gregory, I. Lavender, J. M. Rawson and B. K. Tanner,
J. Chem. Soc., Chem. Commun., 1995, 679–680.
11 C. Allen, D. A. Haynes, C. M. Pask and J. M. Rawson,
CrystEngComm, 2009, 11, 2048–2050.
3
4
A. Alberola, R. J. Less, C. M. Pask, J. M. Rawson, F. Palacio,
P. Oliete, C. Paulsen, A. Yamaguchi, R. D. Farley and
D. M. Murphy, Angew. Chem., Int. Ed., 2003, 42, 4782–4785.
H.-U. Höfs, J. W. Bats, R. Gleiter, G. Hartmann, R. Mews,
M. Eckert-Maksić, H. Oberhammer and G. M. Sheldrick,
Chem. Ber., 1985, 118, 3781–3804; A. J. Banister, I. B. Gorrell,
W. Clegg and K. A. Jørgensen, J. Chem. Soc., Dalton Trans.,
12 J. M. Rawson, A. J. Banister and I. Lavender, Adv. Heterocycl.
Chem., 1995, 62, 137–247.
13 A. J. Banister, N. Bricklebank, W. Clegg, M. R. J. Elsegood,
C. I. Gregory, I. Lavender, J. M. Rawson and B. K. Tanner,
1995, private communication to CSD, refcode ZADVAB.
14 Bruker, APEX II, SAINT and SADABS, Bruker AXS Inc.,
Madison, Wisconsin, USA 2009.
1
991, 1105–1109.
15 G. M. Sheldrick, Acta Crystallogr., Sect. A: Found. Crystallogr.,
2008, 64, 112–122.
5
See for example P. J. Alonso, G. Antonerra, J. I. Martínez,
J. J. Novoa, F. Palacio, J. M. Rawson and J. N. B. Smith,
Appl. Magn. Reson., 2001, 20, 231–247; R. T. Boeré and
T. L. Roemmele, Coord. Chem. Rev., 2000, 210, 369–445.
D. A. Haynes, CrystEngComm, 2011, 13, 4793–4805.
S. A. Fairhurst, K. M. Johnson, L. H. Sutcliffe, K. F. Preston,
A. J. Banister, Z. V. Hauptman and J. Passmore, J. Chem.
Soc., Dalton Trans., 1986, 1465–1472.
16 L. J. Barbour, J. Supramol. Chem., 2001, 1, 189–191.
17 C. F. Macrae, I. J. Bruno, J. A. Chisholm, P. R. Edgington,
P. McCabe, E. Pidcock, L. Rodriguez-Monge, R. Taylor, J. van
de Streek and P. A. Wood, J. Appl. Crystallogr., 2008, 41,
466–470.
6
7
18 B. V. Almelo, X'Pert HighScore Plus., PANalytical, The
Netherlands. 2009.
8
K. V. Shuvaev, A. Decken, F. Grein, T. S. Abedin,
L. K. Thompson and J. Passmore, Dalton Trans.,
19 Jaguar, version 7.7, Schrödinger, LLC, New York, NY, 2010.
20 C. S. Clarke, D. A. Haynes, J. N. B. Smith, A. S. Batsanov,
J. A. K. Howard and J. M. Rawson, CrystEngComm, 2010, 12,
172–185.
21 H. Du, R. C. Haddon, I. Crossing, J. Passmore, J. M. Rawson
and M. J. Schriver, Chem. Commun., 2002, 1836.
2
008, 4029–4037; A. Decken, T. S. Cameron, J. Passmore,
J. M. Rautiainen, R. W. Reed, K. V. Shuvaev and
L. K. Thompson, Inorg. Chem., 2007, 46, 7436–7457; the
data in the following paper gives ΔHdim = −16 kJ mol−1:
This journal is © The Royal Society of Chemistry 2013
CrystEngComm